ABSTRACT
We compare patterns of ocean and coastal 'biodiversity' from analysis of a presence / absence dataset for species of asteroids encompassing the whole Atlantic Ocean (partitioned into 26 regions). Species richness per area in both the shelf (< 200 m) and upper bathyal waters (200-499 m) is more than ten times greater than in the deep ocean (> 500 m). Even total species counts (unadjusted for area) indicate that diversity of coastal areas and the deep sea are equivalent as is taxonomic distinctness, an index which is markedly less sample-size dependent than other common diversity measures. Our results may support the contention of Gray (1994; in press), which questions the emerging paradigm that the deep sea has exceptionally high diversity. It is evident, however, that comparisons (e.g. between coasts and the deep sea) are problematic and can depend very much on the element(s) of biodiversity measured and on sampling methods. Any wider conclusions drawn are therefore only tentative, particularly since assessment is made of only one faunal group.
Highest
species richness per area is found in subtropical latitudes (20-30 degrees
N) of western coastal areas, the region of the Atlantic associated with
greatest coral reef development, although these diversity 'hotspots' may
be a feature of deeper waters as well. The depth range of asteroids was
also found to be significantly correlated with their geographical range.
Distinctive asteroid assemblages are most evident in shelf waters, where six groupings cluster recognisably according to geographical location. Three of these are the southernmost regions of the Atlantic and are characterised by their isolation, long evolutionary history and high levels of endemism.
The usefulness and shortcomings of binary datasets are reviewed. It is concluded that within limitations they can be of value for determining broad (e.g. regional, ocean-scale and global-scale) patterns of diversity and community structure, especially when analysed using measures relatively uninfluenced by sample size.